Musings of a Philadelphia Physician who has served the community for six decades

Return to Home


Computers, Websites, and other Digital Gadgetry
What is novel today is old-hat tomorrow; but what is old-hat to someone today is still novel for someone else. These are our own thoughts about a variety of electronic novelties, for whoever finds them of interest.

Website Development

The website technology supporting Philadelphia Reflections is PHP, MySQL and DHTML. The web hosting service is Internet Planners. The development of this website has provided an opportunity to learn new technology, to try out different techniques for getting noticed by the search engines and the trials and tribulations of dealing with malicious hackers and spammers who range from the annoying to the abusive. This collection of articles documents some of our experiences and we hope that people surfing the web looking for solutions to problems we've encountered will benefit.

The primary purpose of this website is to deliver high quality content on the subjects of Philadelphia, Philadelphia History, medicine, medical economics and other subjects of interest to its author, Dr. George R. Fisher.

However, early in 2006 the site was attacked by spammers who broke in using security holes in the previous implementation of PHP. In the subsequent reconstruction of the site, there's been an opportunity to try out lots of new technology and techniques, some of which are detailed here.

Website Statistics

Today's Philadelphia Reflections was born in June 2006. It had a prior incarnation but it was hacked by Nigerian spammers who took it over and turned it into an email factory.

We scrubbed everything down and rebuilt from scratch, implementing as many PHP and MySQL security features as we could find.

We have done all of the standard things to improve our search engine standings but we are really at a loss to explain the inflection points that can be seen in the graphs.

  • We make an effort to produce clean XHTML 1.1 or HTML 4.01 depending on the user's browser, for both the tags and content
  • We have keywords, description and other relevant meta tags for every page
  • The content is rich, varied, relevant and frequently updated
  • The website structure is simple with complete linkages between pages
  • A current Google sitemap is maintained programatically along with robots.txt, RSS, ATOM and a few other more obscure syndication file types
  • Both the sitemaps, etc. and the URLs themselves have been submitted to the search engines
  • We use static URLs, translated by the Apache mod_rewrite function into dynamic
  • These and other techniques we've picked up along the way are described in the topic "Website Development"

Our home page has a Google Page Rank of 5/10 and the pages vary as follows (as of December 2008):

  • No Info 36%
  • 0    2%
  • 1   28%
  • 2   23%
  • 3    9%
  • 4    1%
  • 5    1%

Google Images is by far the largest source of referrals but we also have many visitors who come to us via the search engines and who like what they see and come back; we would like to express our appreciation to all of our visitors.

{website total visitor statistics}

The dips in the Unique Visitors graph were the result of problems with our ISP ... once they were simply off the air and twice they made software changes without notification or testing.

{website returning visitor statistics}

Floating Three-Column CSS Layout

A current fad in web page styling is to use CSS exclusively to define the basic page sections. The "old" way of doing this was to use tables, but that's no longer stylish. Instead, we are exhorted to use CSS exclusively.

A very common page layout has a head and a foot with three columns sandwiched in between. Philadelphia Reflections uses this layout.

Most descriptions of this layout style that I have found Googling around the Internet involve absolute positioning which very often does not adapt well to differing screen sizes and browser window sizes. What we use here makes use of floating columns, which re-size themselves very nicely.

Several anomalies and quirks should be noted:

  • Each element is defined as a DIV
  • The left, right and center DIVs must be enclosed in a "wrapper" DIV
  • The three columns must be followed by a clear:both DIV
  • The center column must be below the left and right columns
  • The center column actually is as wide as the whole page (try including border-style:solid)

These quirks and anaomalies make me think that maybe this either isn't quite kosher or else may be superceded by later CSS definitions. But for the time being, this works very happily and both the HTML and the CSS validate perfectly well.

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01//EN">
    <title> Floating Three-column CSS example</title>

    <meta http-equiv="Content-Type" content="text/html; charset=ISO-8859-1">

<style type="text/css">

 #head {

 #wrap {

 #left {
  width: 30%;

 #right {

 #center {

 #clear {

 #foot {



    <div id="head">

    <div id="wrap">

      <div id="left">

      <div id="right">

      <div id="center">

      <div id="clear"></div>


    <div id="foot">


DHTML, PHP and MySQL References

Start with

DHTML and CSS for the World Wide Web:
Visual QuickStart Guide
by Jason Cranford Teague

DHTML = HTML + CSS + DOM + JavaScript

This book will get you up and running quickly with client-side programming.

Then you need to learn server-side programming.

PHP is an open source server-side scripting language that is easy to learn and very powerful. MySQL is the same ... open source relational database.

The text book for these technologies is

PHP and MySQL for Dynamic Web Sites:
Visual QuickPro Guide
by Larry Ullman

Master those two books and you'll be creating very powerful scripts on both the client and server side that produce dynamic and elegant results.

While you're in the process of doing this, you will constantly need to reference manuals for syntax, functions, etc. There are many, but two will suffice for 90% of what you need:

PHP Manual

W3 Schools


The markup language used by web browsers continues to evolve. The most current version (as of April 2009) is XHTML 1.1, an XML version of HTML.

Many browsers, most particularly IE, do not support XHTML. Technically speaking, they support only the "text/html" mime type, not "application/xhtml+xml". Lots of web developers have gone to the trouble of sticking closing tags ( />) in their BR, HR, META and INPUT tags and a DOCTYPE at the top but then serve the code as "text/html".

This produces a syntactic mish mash which may be worse than using strict HTML 4.01.

Why "worse"? Because of the possibility of unintended results from providing incorrect instructions to the browser. If you care about the output produced by the browser, which most developers and content providers emphatically do, then you have to be careful about what instructions you give the browser. You simply cannot count on getting what you want if what you're telling the browser to do is syntactically incorrect.

However, it's a little difficult to see just what good XHTML is:

  • There are rumors that it renders the non-image portion of a page as much as 50% faster than HTML, but what with gzip and broadband being pretty common these days, it's hard to see that as an especially compelling reason to be bothered.
  • Furthermore, those browsers that do render XHTML (Mozilla, Firefox) are very picky about syntax and blow up much too easily.
  • And the claim that XHTML is the way to get your web pages onto cell phones and toaster ovens leaves me cold. It's just not believable that the format required for these special devices will be the same as for a computer monitor. (For the current status of handheld support on this site, see How to detect an iPhone and other mobile devices

Internet cognoscenti speak disparagingly of "tag soup" but the Internet is a lot more about content than it is about syntax, so who really cares?

Well, somehow, I do. A little. Since we use PHP on this site, we have the opportunity to figure out what features are supported by a browser and render the correct types of tags, mime-types, etc.

Check out the HTTP headers and the page source to see the following script in action:

  1. It renders XHTML 1.1 whenever it encounters a browser that can support it
  2. It uses output buffering (which demonstrably if illogically improves rendering response time)
  3. It sends the whole thing using gzip compression if the browser will support it
  4. But also, it concedes certain issues based on experience for the sake of a smoothly-operating website
//  This script figures out what kind of mime type (HTML vs XHTML) the browser supports and sends the correct headers
//  It also initiates compression, specifies cache-ing and sends other <meta http-equiv headers
//  My thanks to for the basic idea and structure
//  $_SERVER["ACCEPT"] describes the mime_types a browser supports in a comma-separated list:
//    mime_type,mime_type,mime_type
//  If a browser prefers one mime_type or group of mime_types, it adds a q-value
//    mime_type,mime_type;q=x.x, mime_type,mime_type,mime_type,...,mime_type;q=x.x
//  The q-value is a number between 0.0 and 1.0 ... the higher the number, the greater the preference
//  The idea is that if we can serve more than one mime_type we should serve the browser's higher preference
//  ob_start("ob_gzhandler"); does all the work to compress the output if the browser can handle it
//  ob_start("fix_code"); calls the "fix_code" function instead, so initiating gzip is my responsibility
//  $_SERVER["HTTP_USER_AGENT"] is an opaque decription of the browser itself
//  $_SERVER['HTTP_ACCEPT_ENCODING'] describes compression capabilities
//  I output these three variables as an HTML comment so I can debug things more easily
//  Despite my desire to do things "right", you will see I accomodate myself to the reality of user-supplied content 
//  and browser peculiarities in order to have a working website

function fix_code($buffer)
  # Called for HTML browsers to delete all the lovely close-brackets
  # it's up to me to initiate the gzipping because ob_start is called by "fix_code" instead of "ob_gzhandler"
  if (stristr($_SERVER['HTTP_ACCEPT_ENCODING'], 'gzip'))
    header("Content-Encoding: gzip"); // notifies the far-end to un-gzip 
    return (gzencode(str_replace(" />", ">", $buffer),6,FORCE_GZIP));
      return (str_replace(" />", ">", $buffer));

# default values
$charset          = "UTF-8";       # See
$mime             = "text/html";   # Plain vanilla
$cache_control    = "max-age=200"; # Cache expires after 200 seconds

$xhtml_q          = 0;
$html_q           = 0;

# see
$DOCTYPE_xhtml11  = "<!DOCTYPE html PUBLIC '-//W3C//DTD XHTML 1.1//EN' ''>\n"; 
$DOCTYPE_xhtml10  = "<!DOCTYPE html PUBLIC '-//W3C//DTD XHTML 1.0 Strict//EN' ''>\n";
$DOCTYPE_wap      = "<!DOCTYPE html PUBLIC '-//WAPFORUM//DTD XHTML Mobile 1.2//EN' ''>\n";
$DOCTYPE_html401  = "<!DOCTYPE html PUBLIC '-//W3C//DTD HTML 4.01//EN' ''>\n";
$DOCTYPE_html401l = "<!DOCTYPE html PUBLIC '-//W3C//DTD HTML 4.01 Transitional//EN' ''>\n";

$html_xhtml       = "<html xmlns='' xml:lang='en'>\n\n";
$html_iphone      = "<html xmlns='' xml:lang='en' manifest='iphone.manifest'>\n\n";
$html_html401     = "<html lang='en'>\n\n";
$html_html401_IE  = "<html lang='en' xmlns:v='urn:schemas-microsoft-com:vml'>\n\n";  # xmlns:v='urn:schemas-microsoft-com:vml' is recommended by Google for maps display using IE
$html_plain       = "<html>\n\n";

# parental control tag
$pics_Label       = '(pics-1.1 "" l 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0! v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0) 
	gen true for "" r (n 0 s 0 v 0 l 0 oa 0 ob 0 oc 0 od 0 oe 0 of 0 og 0 oh 0 c 0))';

# I include the following HTML comment for my ongoing debugging purposes

# note that I eval $prolog_type below so that the xml header (if any) gets the right charset
$prolog_type      = '$DOCTYPE_html401l $html_plain $show_info';

# the logic

# W3C Validator
if (stristr($_SERVER["HTTP_USER_AGENT"],"W3C_Validator")) 
  $mime        = "application/xhtml+xml";
    # UTF-8 produces character-type errors
    $charset     = "iso-8859-1";
  $prolog_type = '$xml_header $DOCTYPE_xhtml11 $html_xhtml $show_info';
    # fancy wap-enabled handheld device
        # per and application/xhtml+xml is preferred
//      $mime        = "application/vnd.wap.xhtml+xml";
        $mime        = "application/xhtml+xml";
      $prolog_type = '$xml_header $DOCTYPE_wap $html_plain $show_info';
        # non-wap xhtml-enabled browser
          # retrieve the q values for "application/xhtml+xml" and "text/html"

          if (preg_match('%application/xhtml\+xml[^;]*?;q=([1|0]\.[1-9]+)%i', $_SERVER["HTTP_ACCEPT"], $matches)) 
            $xhtml_q = (float)$matches[1];

          if (preg_match('%text/html[^;]*?;q=([1|0]\.[1-9]+)%i', $_SERVER["HTTP_ACCEPT"], $matches)) 
            $html_q = (float)$matches[1];

          # if the q value for HTML is greater than for XHTML
          # then treat output as HTML 4.01 strict (Opera 9.64, for instance)

          if($html_q > $xhtml_q) 
            $mime        = "text/html";
              # UTF-8 produces character-type errors
              $charset     = "iso-8859-1";
            $prolog_type = '$DOCTYPE_html401 $html_html401 $show_info';

            # otherwise, go with XHTML
                # for the time-being application/xhtml+xml is too strict for us: unless your tags are PERFECT, it blows up
//              $mime        = "application/xhtml+xml";
                $mime        = "text/html";
                # UTF-8 produces character-type errors
                $charset = "iso-8859-1";

              # see "Safari Web Content Guide for iPhone OS" for cache manifest description
              if (stristr($_SERVER["HTTP_USER_AGENT"],"iPhone")) 
                $prolog_type = '$xml_header $DOCTYPE_xhtml11  $html_iphone $show_info';
                  $prolog_type = '$xml_header $DOCTYPE_xhtml11  $html_xhtml $show_info';
            # plain text/html browser
              $mime        = "text/html";
                # UTF-8 produces character-type errors
                $charset     = "iso-8859-1";
              $prolog_type = '$DOCTYPE_html401 $html_html401 $show_info';
                # if the browser doesn't specify any X/HTML mime type, treat like HTML 4.01 Transitional (IE 7, for instance)
                $mime        = "text/html";
                  # UTF-8 produces character-type errors
                  $charset     = "iso-8859-1";
                $prolog_type = '$DOCTYPE_html401l $html_plain $show_info';
                # if IE then include Google's recommended "xmlns:v  ..." 
                  $prolog_type = '$DOCTYPE_html401l $html_html401_IE $show_info';

# output the mime type, prolog type and other <meta http-equiv= variables
header("Content-Type: $mime; charset=$charset");
header("Content-Language: en-us");
header("Vary: Accept");

header("Cache-Control: $cache_control");

header("Content-Script-Type: text/javascript");
header("Content-Style-Type: text/css");
header("imagetoolbar: no");

// parental controls from
header("pics-Label: $pics_Label");

// privacy header created at
header("P3P: policyref=\"\", CP=\"NID DSP NOI COR\"");

$xml_header       = "<?xml version='1.0' encoding='$charset' ?>\n";
eval("\$prolog_type = \"$prolog_type\";");

print $prolog_type;

Here's an interesting article on Doctype Switching:

The Philadelphia Reflections webmaster: George IV

(my thanks to for HTML encoding)

Web Standards Validation

There are two primary aspects of a website that need validation:

1. (X)HTML

You can use the W3C's QA Markup Validation Service.
The URL to test the main page of Philadelphia Reflections is

Firefox has several useful add-ons for (X)HTML validation; one that uses Tidy is here: Html Validator

2. CSS

The W3C has a validation service for CSS, too.
For Philadelphia Reflections, the following URL checks all the CSS definitions in the main page: (note: this validator is a little flakey: it produces different answers for the same file; you have to refresh a couple of times to get the whole story)

Firefox has several useful web developer add-on tools; try this one: Web Developer

Once you've gotten the HTML and CSS basics under control, there are other aspects of your site that you will want to validate:

Broken Links

The W3C will check all your links for both response time and validity.


There is an absolutely lovely program called HTML Tidy, origianlly written by Dave Raggett and decribed by the W3C here:

Calls to Tidy are available in some newer renditions of PHP (sadly, not the one we are using), however, on Widows (only) versions of Firefox and Mozilla, you can download an extension that will provide all the Tidy functions in your browser! ... This a fantastic feature that I use all the time.

Syndication XML Validation

Validating RSS and Atom files is greatly facilitated by It has a number of quirks, the worst of which is that it has a length limitation that we exceed and so we have to provide "short" syndication files since all the feed aggregators use this facility and reject any feeds that aren't validated by it.

Google Sitemap Validation

If you submit a sitemap to Google through their Webmaster Tools facility they will validate your sitemap when they load it. An external validation tool is available here: Validome Google Sitemap(s) Validator

Yahoo and Microsoft have agreed to support Google's Sitemap protocol and to support the inclusion of the line "Sitemap:" in robots.txt. If other search engines adopt this facility it will make it much easier to get into the world's many search engines ... they'll pick up this line instead of us having to hunt them down.

Meta Tag Validator

As you puzzle the mysteries of search-engine indexing, you'll want to check your meta tags:

gzip Compression & Headers

When you start getting really fancy and want to include automatic gzip compression, you'll want to see it in action and you'll want to check out all of your HTTP headers:

Response Time

Of course, the reason you''re experimenting with gzip is because you're concerned about response time.
(1) Try this site for a detailed analysis:
(2) Firefox to the rescue again: FasterFox is another lovely add-in:

Geo Tags

Check the validity of your geo tags here: Geo-Tag Validator

Big List is the mother of all lists of validations routines

CSS Zen Garden Suggestions

Here is a list of links (that open in their own pages) that show some of my favorite web designs. The CSS Zen Garden is a website that illustrates what can be done with clever CSS design. The HTML and the content are exactly the same in each of these links, only the CSS changes; but what a difference!

C Note

Dark Rose

Dead or Alive

Egyptian Dawn






Zen Pool


White Lily

Another website to consider is

PHP script to display Google PageRank

Like so many things on this website, the code to find the Google PageRank of the pages was lifted from someone else's work. This work is particularly praiseworthy because it worked exactly as described the minute I got it implemented.

See PHP script to display Google PageRank


define('GOOGLE_MAGIC', 0xE6359A60);
class pageRank{
var $pr; 
 function zeroFill($a, $b){
 $z = hexdec(80000000);
  if ($z & $a){
   $a = ($a>>1);
   $a &= (~$z);
   $a |= 0x40000000;
   $a = ($a>>($b-1));
   $a = ($a>>$b);
 return $a;
 function mix($a,$b,$c) {
   $a -= $b; $a -= $c; $a ^= ($this->zeroFill($c,13));
   $b -= $c; $b -= $a; $b ^= ($a<<8);
   $c -= $a; $c -= $b; $c ^= ($this->zeroFill($b,13));
   $a -= $b; $a -= $c; $a ^= ($this->zeroFill($c,12));
   $b -= $c; $b -= $a; $b ^= ($a<<16);
   $c -= $a; $c -= $b; $c ^= ($this->zeroFill($b,5));
   $a -= $b; $a -= $c; $a ^= ($this->zeroFill($c,3));
   $b -= $c; $b -= $a; $b ^= ($a<<10);
   $c -= $a; $c -= $b; $c ^= ($this->zeroFill($b,15));
   return array($a,$b,$c);
 function GoogleCH($url, $length=null, $init=GOOGLE_MAGIC) {
  if(is_null($length)) {
   $length = sizeof($url);
  $a = $b = 0x9E3779B9;
  $c = $init;
  $k = 0;
  $len = $length;
  while($len >= 12) {
   $a += ($url[$k+0] +($url[$k+1]<<8) +($url[$k+2]<<16) +($url[$k+3]<<24));
   $b += ($url[$k+4] +($url[$k+5]<<8) +($url[$k+6]<<16) +($url[$k+7]<<24));
   $c += ($url[$k+8] +($url[$k+9]<<8) +($url[$k+10]<<16)+($url[$k+11]<<24));
   $mix = $this->mix($a,$b,$c);
   $a = $mix[0]; $b = $mix[1]; $c = $mix[2];
   $k += 12;
   $len -= 12;
  $c += $length;
   case 11: $c+=($url[$k+10]<<24);
   case 10: $c+=($url[$k+9]<<16);
   case 9 : $c+=($url[$k+8]<<8);
   /* the first byte of c is reserved for the length */
   case 8 : $b+=($url[$k+7]<<24);
   case 7 : $b+=($url[$k+6]<<16);
   case 6 : $b+=($url[$k+5]<<8);
   case 5 : $b+=($url[$k+4]);
   case 4 : $a+=($url[$k+3]<<24);
   case 3 : $a+=($url[$k+2]<<16);
   case 2 : $a+=($url[$k+1]<<8);
   case 1 : $a+=($url[$k+0]);
  $mix = $this->mix($a,$b,$c);
 /* report the result */
 return $mix[2];
 //converts a string into an array of integers containing the numeric value of the char
 function strord($string) {
  for($i=0;$i<strlen($string);$i++) {
   $result[$i] = ord($string{$i});
 return $result;
 function printrank($url){
  $ch = "6".$this->GoogleCH($this->strord("info:" . $url));
  $fp = fsockopen("", 80, $errno, $errstr, 30);
  if (!$fp) {
     echo "$errstr ($errno)<br />\n";
  } else {
     $out = "GET /search?client=navclient-auto&ch=" . $ch .  

"&features=Rank&q=info:" . $url . " HTTP/1.1\r\n" ;
     $out .= "Host:\r\n" ;
     $out .= "Connection: Close\r\n\r\n" ; 
     fwrite($fp, $out);
     while (!feof($fp)) {
       $data = fgets($fp, 128);
       $pos = strpos($data, "Rank_");
         if($pos === false){
           $pagerank = substr($data, $pos + 9);

// Display pagerank image. Create your own or download images I made for this script. 
// If you make your own make sure to call them pr0.gif, pr1.gif, pr2.gif etc.

 function pr_image($pagerank){
  if($pagerank == 0){
   $this->pr = "<img src=\"images/pr0.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 1){
   $this->pr = "<img src=\"images/pr1.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 2){
   $this->pr = "<img src=\"images/pr2.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 3){
   $this->pr = "<img src=\"images/pr3.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 4){
   $this->pr = "<img src=\"images/pr4.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 5){
   $this->pr = "<img src=\"images/pr5.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 6){
   $this->pr = "<img src=\"images/pr6.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 7){
   $this->pr = "<img src=\"images/pr7.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 8){
   $this->pr = "<img src=\"images/pr8.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   }elseif($pagerank == 9){
   $this->pr = "<img src=\"images/pr9.gif\" alt=\"PageRank " .$pagerank. " 

out of 10\">" ;
   $this->pr = "<img src=\"images/pr10.gif\" alt=\"PageRank " .$pagerank. 

" out of 10\">" ;
 function get_pr(){
  return $this->pr;


Do following:

   1. Save the code above as pagerank.php.
   2. Download or create your own images to display each rank.
   3. Create a directory "images" containing all page rank images. 
   4. See code below on how to use the class. 

$gpr = new pageRank();
//display image
echo $gpr->get_pr();

(my thanks to for HTML encoding)

Regex URL Matching

On this site we check for the existence of a URL whenever an entry is updated

There are two key technologies at work

  • A PHP function that checks whether a URL is valid (thanks to marufit at gmail dot com in the PHP Manual)

  • Regex (regular expression) in a preg_replace_callback routine; this one is mine, all mine

function url_exists($url) 
// checks whether a URL actually exists on the Internet
$handle   = curl_init($url);
if (false === $handle)
    return false;
curl_setopt($handle, CURLOPT_HEADER, false);
curl_setopt($handle, CURLOPT_FAILONERROR, true); 
curl_setopt($handle, CURLOPT_NOBODY, true);
curl_setopt($handle, CURLOPT_RETURNTRANSFER, false);
$connectable = curl_exec($handle);
return $connectable;

function aExists($matches)
// function called by preg_replace_callback
// $matches[0] is the complete match
// $matches[1] the match for the first subpattern
//	enclosed in '(...)' and so on

// checks to see if a regular link exists
// something similar is done for img src= also

$srcURL = $matches[3];
if (url_exists($srcURL)) {do something; return "";}  
else {do something else; return "";}

$foo = preg_replace_callback(
            '/(.*?)(<a .*?href=")([^"]*)("[^>]*>)(.*?)(<\/a>)/i',

(my thanks to for HTML encoding)

Parsing name-value pair attributes in an HTML tag

Not only do the attributes in an HTML tag come in random order but many are optional

Here's a regex solution:

function tagAttr($matches) {print_r($matches);}

$string = '<img src="/images/picture.jpg" width="300" class="left" alt="alt keywords" />';

$foo	= preg_replace_callback(

Produces the following:

    [0] => <img src="/images/picture.jpg" width="300" class="left" alt="alt keywords" />
    [1] => alt keywords
    [2] => left
    [3] => 
    [4] => /images/picture.jpg
    [5] => 
    [6] => 300

The regex is a series of alternating sequences; so, add href="([^"]*)"| in front of alt="([^"]*)" to select an additional attribute.

$matches[0] is the complete match
$matches[1] is alt=
$matches[2] is class=
$matches[3] is style=
$matches[4] is src=
$matches[5] is height=
$matches[6] is width=

My thanks (a) to Flagrant Badassery for putting me onto the idea and (b) to for HTML encoding

Regular Expressions

Anyone who has used the expression *.doc to search for Word files has used Regular Expressions ("regex") without realizing it. Regex arose from mathematical theory and is available in many programming languages; it is simply the only way to deal with large amounts of text. And yet most people are completely unaware of it.

Philadelphia Reflections uses regex extensively for two primary purposes: (1) checking input from forms and (2) modifying HTML input in during the creation of articles for the site.

The text PHP and MySQL by Larry Ullman has a very good introduction to regex in his chapter on security.

The great advantage of regex is that it can identify very complex patterns in a mass of text. The great disadvantage of regex is that it has developed in sort of an underground way and there exist numerous varieties that are essentially incompatible. PHP offers two regex functions: one for the POSIX Extended variety of regex and he other for the Perl language compatible vesion called PCRE. POSIX is less powerful but far easier to learn. JavaScript offers its own variety of regex which isn't quite the same as either of the two PHP versions.

References include the Ullman book, the PHP online manual has a number of handy tips on regex use in its two supported varieties, the O'Reilly book Mastering Regular Expressions is interesting and Jan Goyvaerts has a very helpful website ( and book Regular Expressions: The Complete Tutorial.

My experience is that this area requires diligent hacking which may be sub optimal but unavoidable ... for this purpose, Jan Goyvaerts' Regex Buddy is indispensible; you simply must get this program if you hope to make anything of Regex.

Here are examples of checking for a valid email address in both Javascript and PHP:


// check email

var namePattern = /^[a-zA-Z0-9][a-z0-9_.-]*@[a-z0-9.-]+\.[a-z]{2,4}$/i; = trim(;
if	(
	( > 0)
	(! == "[none]")
			alert("Please enter a valid email address");;
			problem = "yes";
       			return false;


// check email

$emailpattern = "^[a-zA-Z0-9][a-z0-9_.-]*@[a-z0-9.-]+\.[a-z]{2,4}$";
if	(
	(trim(strlen($_POST['email']))  > 0)
	(!$_POST['email'] == "[none]")

	(!eregi ($emailpattern, stripslashes(trim($_POST['email']))))
		$inputerror		=	TRUE;
		$inputerrormessage	.=	"<br />* An invalid email address was entered";

Incomprehensible? Yes, absolutely.

Useful? More than you can realize until you are actually faced with the problem of, say, verifying that a user has input a valid email format, or trying to figure out whether a user-input IMG tag is using the correct syntax; or else maybe trying to convert a huge web page from XHTML 1.1 to HTML 4.01 because you've determined that the browser is syntactically crippled.

And, once you get deep into it, the stuff is actually intriguing and fun.

Process .htm and .html as php

It is sometimes helpful to include php scripting in files that do not have the file extension of php.

There is quite a lot of discussion on the web about this but at least on this server the answer is not what most people think.

In the .htaccess file in the root folder include these two lines:

AddType application/x-httpd-php .html .php .htm
AddHandler application/x-httpd-php .html .php

Editing HTML with PHP scripts

When creating scripts that allow a user to edit HTML, you have to ensure that the browser doesn't confuse the input with HTML to be rendered. I struggled with this long and hard and throughout the utilities section of this website are various hacks that I created with brute force. They work, but they are mostly ugly and all were time consuming.

Well, guess what? The PHP manual has a section on this subject and the solution is really rather elegant. Chaper 56. PHP and HTML. It's worth reading, but the essential bits are reproduced below:

Example 56-1. A hidden HTML form element
   echo "<input type='hidden' value='" . htmlspecialchars($data) . "' />\n";

Example 56-2. Data to be edited by the user
   echo "<textarea name='mydata'>\n";
   echo htmlspecialchars($data)."\n";
   echo "</textarea>";

Example 56-3. In a URL
   echo "<a href='" . htmlspecialchars("/nextpage.php?stage=23&data=" .
       urlencode($data)) . "'>\n";

Static vs Dynamic URLs

It used to be that no spiders or search engines could index a dynamic URL, namely one that contained a "?" followed by parameters to be used by PHP, ASP or other server-side scripting languages to drive a website using a database.

Nowadays, Google and Yahoo seem to do a perfectly fine job of indexing dynamic URLs but Google has a disclaimer warning that it may still encounter problems with dynamic URLs and the SEO literature is still full of warnings that other spiders and search engines may be blind to everything to the right of the "?".

Furthermore, a *.php extension is an invitation to bad guys to try to break in and wreak many sorts of havoc: this site was hacked by Nigerians a few years ago using PHP tricks and they managed to use it as an email factory until our ISP shut us down. I came on the scene at that point and implemented every safeguard I could find, but the concern still lingers.

Finally, dynamic URLs are not user friendly ... human beings generally do not know what to make of long strings of obscure parameters.

Apache has a feature called "mod_rewrite" that allows you to specify, via regex, that you want incoming URLs to be transformed in some way. Apache's instructions on this subject are here: URL Rewriting Guide. I have used that facility at Philadelphia Reflections to use static URLs for public use while still allowing me to use parameters to drive the website with the database.

Two excellent articles on this got me started:

Here's what I did:

Step 1: htaccess

I added these lines to the htaccess file

Options +FollowSymLinks
RewriteEngine on
RewriteRule ^(blog|topic)/([0-9]+)\.html?$ reflections.php?type=$1&key=$2
  1. ^(blog|topic)/([0-9]+)\.html?$ is the pattern to be compared against all incoming URLs.

    If matched, it changes the URL to reflections.php?type=$1&key=$2

  2. The ^ ... $ sequence in the pattern says that we will match the whole string, not just some part in the middle
  3. (blog|topic)/ matches either   "blog"   or   "topic"   followed by   "/"  
  4. ([0-9]+) matches one or more digits
  5. \.htm matches   ".htm"  
  6. l? matches 0 or 1 lower-case Ls (so that we will match either htm or html)
  7. In the replacement string $1 is replaced with the contents of the first () in the pattern: either "blog" or "topic"
    ... and $2 is replaced by the second () in the pattern, namely the numeric ID on the database of the blog or topic

The result is that

is transformed into

The latter is what is passed in to me in the reflections.php routine, which tells me to pull up blog #906 from the database.

Both of these URLs are equivalent to the old, ugly dynamic URL

which still works, in case there are any legacy bookmarks or links out there, but going forward the new, simple, static URL is the face we will present to the world.

Step 2. SMOP

After the htaccess regex was debugged, all that was left was a simple matter of programming. In fact, I had to completely rewrite the driver script, reflections.php, and the XML creation script which creates the RSS, sitemap, etc. files; plus a lot more besides. It was a lot of work but the breakthrough was in figuring out the htaccess trick; everything else was just work.

In July 2008, after Volumes were implemented, another RewriteRule was implemented:

RewriteRule ^volumes?/([0-9]+)\.html?$ volume.php?table_key=$1

Ampersand Madness: Convert & to &amp; to prevent XHTML errors

The whole subject of "encoding" gives me a headache.

Encoding In General

The first thing you have to know is: what is HTML encoding ... so look here:
or here:

(These are HTML encodings; URL encoding is something else again ... look here:

Ampersand Encoding and Conversion

Later on, you'll find out that the ampersand is a huge source of XHTML errors because it has to be written

  • &amp;
  • &#38;
  • &#x26;

but you will struggle endlessly with how to get the darn thing to stay converted. First of all, content providers feel justifiably justified in including bare naked "&"s wherever they please; second of all, you will find that encoded ampersands get stripped back to their bare naked selves by browsers and other well-meaning sorts.

So, my undying thanks to Michael Ash's Regex Blog for providing the regex pattern in the following bit of PHP code:

$pattern = '/&(?!(?i:\#((x([\dA-F]){1,5})|(104857[0-5]|10485[0-6]\d|1048[0-4]\d\d|104[0-7]\d{3}|10[0-3]\d{4}|0?\d{1,6}))|([A-Za-z\d.]{2,31}));)/i';
$replacement = '&amp;';
$string = preg_replace ( $pattern, $replacement, $string);

I don't know how it can possibly work, and I may yet eat my words, but for the moment it seems to do the trick.

Ampersand Encoding In RSS

Another thing: &#x26; is the only ampersand encoding form acceptable to both RSS and Atom. So, look at the souce of this page and you will find that I use this encoding in the title ... that's because the title goes into the Title field of my RSS and Atom feeds.

HTML Forms

How do you (a) open a form when a radio button is clicked (b) in a new window?

Here's how it's done on this website.

  • The radio button is activated by a little JavaScript routine

  • The new window is simply a matter of including the target attribute in the form tag

<script type="text/javascript">

	/* javascript function called by the radio buttons 
	     to submit the form when clicked */

	function formSubmit()


  <form name="form_x" id="form_x"

	<legend>legend surrounding the form</legend>

	<input type="radio" name="key" value="1269" onclick="formSubmit()" />



(my thanks to for HTML encoding)

Geo Positioning

Geo Tagging refers to adding latitude and longitude information to websites and photographs. This has been around for a long time but it has taken the advent of Google Earth for it to really start to catch on.

This blog entry has geo meta tags that you can see if you look at the HTML source ("View > [Page] Source"). The input was as follows:

Address: 82 Devonshire St Boston MA

Lat: 42.3578 Lon: -71.0577

Descriptive Place Name: Fidelity Investments headquarters

Region: US-MA Country Code: US Country Name: United States

This creates meta tags in the HTML Header as follows:

<!-- geo tags for 82 Devonshire St Boston MA -->
<meta name="ICBM"          content="42.3578, -71.0577" />

<meta name=""   content="US" />
<meta name="geo.region"    content="US-MA" />
<meta name="geo.placename" content="Fidelity Investments headquarters" />
<meta name="geo.position"  content="42.3578; -71.0577" />

<meta name=""      content="Fidelity Investments headquarters" />
<meta name="tgn.nation"    content="United States" />
  • To be precise you can go to the location with a GPS device and record the exact coordinates.
    (see below for a discussion of GPX, the GPS transfer protocol)
  • Google Earth is a very good way to find coordinates of any spot on earth
    and both GE itself and the KML Editor will allow you to pick up coordinates from GE.
  • You can find the lat and lon of a US address using GeoURL Header Generator
  • My Geo Position can also be helpful

  • But for most lookups, the easiest method is to use the Firefox Sidebar AddOn called Minimap; it's right in your browser so you don't have to switch back and forth. (The AddOn download is found here: Mini Map Sidebar).

The Region, Country Code and Country Name can be found here: ISO-3166-1 Country Names

geo.placename and are often rendered as the city name but are intended to describe the geographical feature ("Pyramids of Giza" or something). This tag is optional.

HTML geo meta tags can be validated here: {geo tag validation}

There is a search engine of long standing that reads HTML geo meta tags and indexes the website based upon its location; for searching, it groups sites based on their geographic proximity: GeoURL.

Photographs can also contain geo meta data, so-called EXIF data (Firefox has an EXIF viewer AddOn).

JPEG is the most common image format and the easiest to deal with. The combination of Picassa2 and Google Earth allow you easily to add this information to your own photos.

The process of adding lat and lon to your photographs is this:

1. Select one or more photos in Picassa
2. Select Tools > GeoTag > GeoTag with Google Earth ...
3. This starts Google Earth and you can "fly" to the location of the picture

4. A small Picasa window will appear in Earth's lower-right corner displaying thumbnails of the pictures you selected; press the "Geotag" button.
5. When all of your pictures are tagged, press the "Done" button

Slowly, camera manufacturers are providing GPS capability. Some few have GPS devices built in and some others allow an external GPS device to be attached, although both Canon and Nikon are way behind the curve ... if you own either, you can essentially forget it: the best - lousy - solution is to carry around a GPS with you and synchronize the times ... ugly.

The Google Maps API allows maps to be embedded in a website as is done here. Google Maps API

The JavaScript required to embed the map on this page can also be seen in the HTML source ("View > [Page] Source"). In addition to JavaScript, you need a DIV with an ID of "map" or whatever is specified in the JavaScript document.getElementById entry, which specifies the height and width of the map to be displayed.

To embed these maps you must register with Google

In addition, there are extensions to ATOM and RSS to include lat and lon in your syndication feeds; there are three standards that I have found: GeoRSS (ATOM example) , W3C Geo (RSS example) and an "ICBM RSS Module". This website extends the namespaces of both its ATOM feed and its RSS feed to include all the tags.

Google, Yahoo and Microsoft all now support GeoRSS as a feed to their map programs. My sense of it is that KML is a richer protocol, allowing more features, but fundamentally all these XML variants do mostly the same thing.

Google Sitemaps can include links to KML files (and ATOM, now, too). Part of the sitemap generation on this site is some code that picks up every *.kml and *.kmz file in the /kml/ folder and adds them to our sitemap.xml file.

Google Earth is filled with delights, not the least of which is a Flight Simulator! Google Earth Flight Simulator Keyboard Controls

KML ( Keyhole Markup Language, Keyhole being the predecessor to Google Earth) is an XML protocol that allows you to incorporate Google Earth into graphical presentations. Google KML Overview

Google Earth Outreach helps you get started: Google Earth Outreach

An extraordinary collection of KML files you can view is found here: Spectacular satellite images of the world

I found a KML editor here: NorthGates' KML Editor for Windows. It's rudimentary but very handy for what it does do.

Here's the Google Earth tools list where I found the KML editor: EarthPlot Software Tools For Google Earth

The way we serve the KML in the link that connects to Google Earth from individual blogs uses the following PHP script as its base:


// See Google Earth's KML 2.1 Reference

$lat			= $_GET['lat'];
$lon			= $_GET['lon'];
$placename		= $_GET['placename'];
$altitude		= $_GET['altitude'];
$range			= $_GET['range'];
$heading		= $_GET['heading'];
$tilt			= $_GET['tilt'];
if ($altitude	== NULL) {$altitude	= 0;}
if ($range	== NULL) {$range	= 1000;}
if ($heading	== NULL) {$heading	= 0;}
if ($tilt	== NULL) {$tilt		= 0;}
$description	= "<h3><font color=\"#ea9f20\"><a href=\"\">
		<p>The musings of a Philadelphia physician who has served the community for six decades.</p>";
header('Content-Type: application/');
header('Content-Disposition: inline; filename="philadelphia-reflections.kml"');

echo '<?xml version="1.0" encoding="UTF-8"?>'; 


<kml xmlns="">

    <name><?php echo $placename; ?></name>
        <![CDATA[<?php echo $description; ?>]]>
      <longitude><?php echo $lon; ?></longitude>
      <latitude><?php echo $lat; ?></latitude>
      <altitude><?php echo $altitude; ?></altitude>
      <range><?php echo $range; ?></range>
      <tilt><?php echo $tilt; ?></tilt>
      <heading><?php echo $heading; ?></heading>
      <coordinates><?php echo "$lon,$lat,$altitude"; ?></coordinates>


Of course, GPS devices are an integral part of this process of Geo Positioning. GPS devices are supposed to support the open-source protocol GPX,
which is an XML-based description of waypoints and routes. Wikipedia describes GPX here: GPS eXchange Format

The GPX protocol's official website is here: GPX The GPS Exchange Format

Google Earth supports raw GPX (File > Open ...) and when you open a GPX file in Google Earth, it converts it to KML. But if you want stand alone programs to do this:

If you have non-standard GPS data, you may want to have a look at GPS Babel for conversion of native GPS formats as well as the GPS Utility and G7ToWin

A nice blog on these things relative to Google Maps is here: Using XSL to Transform Google Earth (KML) and GPX to Google Maps API

At Philadelphia Reflections, we are creating tours by carrying a GPS and a camera around on our travels. The GPS track becomes a path and waypoints become placemarks. When you come home, download the GPS data in GPX format and open up the GPX file in Google Earth. Use Google Earth to edit the placemark balloons, including pictures and text.

There are many, many sightseeing blogs around that take you to interesting places on Google Maps and Google Earth. A place to start looking is Sightseeing with Google Satellite Maps

Somehow, the concept of "mashup" is related to all of this but it sort of sounds like the term "multimedia" a few years ago ... fancy in concept but somewhat vague in reality.

Google has a Mashup Editor and Wikipedia has a definition but it's not clear what it all adds up to.

(my thanks to for HTML encoding)

Send a KML file from disk using PHP

Sending a kml or kmz disk file is as easy as clicking on it. But different browsers react differently, some asking you which program to use others storing the file on your desk top, etc. Preprocessing the file through PHP can reduce some of these annoyances.


// reads and sends a kml or kmz file
// located in /whatever/kml/
// calling protocol:
// this-program.php?file=somefile.kml

// read the input and check that it's a kmz or kml file
// ....................................................

$kml_file	= $_GET['file'];

if (($kml_file === NULL) or ($kml_file == "")) {exit ("error message");}
if ((substr($kml_file, -4) != ".kmz") AND (substr($kml_file, -4) != ".kml"))	
	exit ("error message");

// prepend the file path information to the file name and check that it exists
// ...........................................................................

$kml_file_name = "/whatever/kml/" . $kml_file;
if (!file_exists($kml_file_name)) { die ("error message");}

// send out the HTTP header information followed by the file contents
// ..................................................................
header("Cache-Control: no-cache, no-store, must-revalidate"); // trying to keep from getting the files stored on the local computer
header("Expires: Mon, 26 Jul 1997 05:00:00 GMT");

if (substr($kml_file, -4) == ".kml") {header('Content-Type: application/');}
if (substr($kml_file, -4) == ".kmz") {header('Content-Type: application/');}

header("Content-Disposition: inline");
header("Content-Description: KML or KMZ data intended for Google Earth");

readfile ($kml_file_name);


Webpage Printing

This site offers a Print button for all Reflections and Topics. Formatting the text on the pages to print nicely works quite well; but how to specify what to do with images remains a bit unclear (as of August 2006). Although 95% of users employ Internet Explorer because Microsoft supplies it free with new computers, IE is just about the worst browser to use for printing. Safari is much better, and Firefox is pretty good. Opera is also satisfactory, but Internet Explorer is not recommended. The other browsers are free; find them in Google and download them. For the usual user, that's all you have to know.

If you are curious about the technicalities, read on. The "trick", if it can be called that, to special print formatting is the media attribute for CSS styling. The main stylesheet for this website is called in a LINK statement as follows:

<link rel="stylesheet" type="text/css" media="all" href="stylesheets/reflectionsLayout.css">

The media attribute tells the browser to use this sylesheet for all media types, i.e., for screens and printers. In the pages that are formattted to print is a stylesheet that cascades below the main stylesheet and therefore supercedes it. This stylesheet controls the printing. IE seems to have its own views on font size so we use some conditional comments to coax it to our way of thinking.

Here and there throughout the website are pages that contain onscreen navigation ("jump to top" and that sort of thing). We hide them when printing by saying class="navstrip" which you can see will result in those elements being hidden.

The specification of

<body onload="window.print()">

(all lower case for XHTML purposes) is what forces the print dialog to appear.

The remaining problem is how to specify CSS formatting for images so that text flows around them as we want. The formatting seems to work on screen for all browsers but only on some browsers for printing.

<style type="text/css" media="print">

  body        { margin: 0; padding: 0; width: 100%; }				
    #wrapper    { margin: 0; padding: 0; width: 100%; }
      #center     { margin: 0; padding: 0; }

      #content    { font-size: 11pt; line-height: 100%; font-family: "Times New Roman", Times, serif; }

        .navstrip   { visibility: hidden; }

<!--[if IE]>
<style type="text/css" media="print">

      #content    { font-size: 14pt; }

(my thanks to for HTML encoding)

Font Families

The following link shows the results of a survey done to find out which font families are installed on Windows machines. This should help determine which fonts to use.

Identifont is a site that helps identify good font choices.

Open a new window with XHTML

Once upon a time you could say

<a href="link" target="_blank">Click to open a new window</a>

and a new window would open. Highly annoying if used very often, but sometimes it's the right thing to do.

And then XHTML comes along and this is not longer legal.

target="_blank" is "deprecated" without a single word as to what a poor developer is to substitute.

Here's what you do:

<a href="link" onclick="; return false;">Click to open a new window</a>

HTML Anchor without an HREF?

Sometimes I want to execute a JavaScript function when a user clicks a link, but nothing else.

If I omit the href entirely, the cursor doesn't change and some browsers don't recognize the text as a link:

<a onclick="function();">

If I include the pound sign, which seems to be a very popular trick, I get sent to the top of the current page, which messes up both History and the backspace button; to say nothing of the fact that I don't want to jump to the top of the page:

<a href="#" onclick="function();">

Including the function in the href and omitting the onclick seems to be the answer to my specific problem:

<a href="javascript: function();">

Once again, my thanks to for HTML encoding.

Create and send CSV files from PHP

Here's how CSV files are created and downloaded on this site. No saving the file or data import into Excel ... Excel just opens with the data automatically. Very handy.

This function as shown pulls all the field names to create a CSV header and then pulls every field from every row in the table. There is no need to know the field names, the data types or the size of the table. Quotes in the data are double-quoted and the result is surrounded by quotes.

CSV calls for each field to be contained in double quotes.

  • For strings that may contain double quotes and need to be escaped, I do this:
    $title = str_replace('"', '""', $title);
  • For these and other strings that may contain special characters and need to be quoted, I do this:
    $csv_output .= "$var1,$var2,\"$title\",$var3\n";

Pull data from a database using standard PHP MySQL functions:


$db_link        = mysql_connect(DB_HOST, DB_USER, DB_PSWD);
$db_selected    = mysql_select_db(DB_DATABASE, $db_link);

# Create the CSV file header from the database-table field names
$query          = "SHOW COLUMNS FROM table";
$result         = mysql_query($query);

$csv_output     = NULL;
while ($row = mysql_fetch_assoc($result))
	$csv_output .= '"' . str_replace('"', '""', $row["Field"]) . '",';
$csv_output  = substr($csv_output, 0, -1) . "\n";  // remove trailing "," and add a line break

# Pull all the rows
$query          = "SELECT * FROM table";
$result         = mysql_query($query);

# loop through database records creating one comma-delimed line per row
while ($row = mysql_fetch_assoc($result))
	foreach ($row as $key => $value)
	  $$key = $value;
	  $$key = str_replace('"', '""', $$key);
	  $var  = $$key;
	  $csv_output .= "\"$var\",";
	$csv_output .= "\n";

# send the file
$size_in_bytes		= strlen($csv_output);
$csv_file		= "filename_" . date("Y-m-d") . ".csv";
$ContentType		= "Content-type: application/";
$ContentLength		= "Content-Length: $size_in_bytes";
$ContentDisposition	= "Content-Disposition: attachment; filename=\"$csv_file\"";


echo "$csv_output"; 


I generally use compression with output buffering to speed things up:


To use in your HTML:

<button onclick="window.location='CSVoutput.php'" 
	style="font-size:85%;width:100px;">Download<br />CSV file</button>

(my thanks to for HTML encoding)

PHP out of memory condition

Fatal error: Allowed memory size of 18388608 bytes exhausted 
(tried to allocate 724 bytes) in /home/dir1/dir2/script.php on line ###

When working with large amounts of data in memory (very long concatenated strings and/or very large arrays in my case), the server may hit a memory max. No amount of "unset" of variables will do the trick past a certain point.

This is the result of a memory allocation ceiling set in php.ini that can be over-ridden (judiciously) as follows:

ini_set  ("memory_limit", -1  );

Be sure to test your code with smaller amounts of data first: this limit is set for a reason ... programs have been known to have been poorly written (not yours, of course; but test anyway).

Javascript: document.write and XHTML

For reasons that make no sense to me, the Javascript command document.write does not work when your page is rendered properly in XHMTL (as described elsewhere in this Topic).

I have searched the web in vain to find a Javascript solution. Many are offered but none work worth a damn.

So don't bother. Use PHP's echo function. It works perfectly.


With the rise of spam entries in web forms, a security feature called "captcha" has been developed.

CAPTCHA stands for "Completely Automated Public Turing test to tell Computers and Humans Apart". The idea is that only a human could read the letters contained in the image and then enter them in the form. "Accessibility", ie., designing websites to accommodate people with handicaps is obviously hindered by Captcha; but at least given our experience with this website, spamming is a huge problem and the inability of handicapped people to leave comments is a price we are willing to pay to rid ourselves of spam. The W3C, the Godhead of web standards, does not agree with me and lectures at length on the futility of captcha: Inaccessibility of CAPTCHA. Whatever. I may get around to implementing some of their recommendations later, if we continue to be spammed.

Spammers have countered captcha in a number of ways. The first is OCR, which is why the images have fuzzy backgrounds and distorted letters: trying to defeat OCR programs. As OCR techniques have improved, captcha programs have moved from letters to "objects" such as kittens, boxes, etc., which are thought to be harder for computers to recognize; harder for people, too: cat vs kitten, for example. I am amazed to learn during my captcha research that there are spammers who offer micro-payments to people in India, etc. to enter hundreds of spam manually in captcha-ed websites that have defeated their automated spamming systems. Move, counter move; seemingly endlessly.

In this website captcha has been implemented using PHP: the comments form that appears at the end of every page has an image created using the PHP image-creation routines which has random characters in it. If the characters in the image are entered correctly in the form, the comments are entered into the database.

I cribbed the PHP captcha code from and it worked right out of the box with the minor exception that the form HTML didn't quite pass XHTML muster; easily fixed. (I have subsequently discovered that PHP security and sessions don't play well together; this problem remains unresolved and I've had to turn off captcha processing for my secure pages.)

I implemented a number of other spam counter measures before I got around to captcha, which involved noticing what the spammers did and writing code to frustrate it. I am constantly on the lookout for new security techniques to implement.

RSS, Atom, Syndication, etc.

The world is full of XML and XML-like file formats for syndication purposes

Here's the list of files we generate automatically for submission to search engines and such.

(For right now, things are a bit abbreviated) (RSS Syndication file) (Atom Syndication file) (Google sitemap) (A9/Amazon siteinfo.xml) (Yahoo IDIF file 1) (Yahoo IDIF file 2) (Yahoo IDIF file 3) (Yahoo IDIF file 4) (Yahoo IDIF file 5) (Yahoo IDIF pointer file) (Yahoo urllist.txt)

Validate Short RSS | The Short RSS File itself
Validate Short rss (lower case) | The Short RSS File itself (lower case)
Validate Short ATOM | The Short ATOM File itself extended successfully pinged successfully pinged successfully pinged
Technorati successfully pinged
Ping-O-Matic successfully pinged
Syndic8 successfully pinged (Feed ID 477463)

Ping Blogroller manually
Ping MyYahoo manually

The RSS and Atom validator ( has a length restriction. I don't know what it is, exactly, but it bombs if your file is "too long". Since most syndication readers run the validator before they'll accept a feed, I have resorted to creating a short file, which is what I point to in my meta tags.

Here's how I provide change frequency and priority for our Google sitemap (in PHP ... $mod is the variable containing the date last modified)

$GOOGLEpriority = "0.0"; $GOOGLEfreq = "yearly";	// default

if ($mod > mktime(0,0,0) - 86400*210)	{$GOOGLEpriority = "0.1"; $GOOGLEfreq = "monthly";}	// past 210 days
if ($mod > mktime(0,0,0) - 86400*180)	{$GOOGLEpriority = "0.2"; $GOOGLEfreq = "monthly";}	// past 180 days
if ($mod > mktime(0,0,0) - 86400*150)	{$GOOGLEpriority = "0.3"; $GOOGLEfreq = "monthly";}	// past 150 days
if ($mod > mktime(0,0,0) - 86400*120)	{$GOOGLEpriority = "0.4"; $GOOGLEfreq = "monthly";}	// past 120 days
if ($mod > mktime(0,0,0) - 86400*90)	{$GOOGLEpriority = "0.5"; $GOOGLEfreq = "monthly";}	// past 90 days
if ($mod > mktime(0,0,0) - 86400*60)	{$GOOGLEpriority = "0.6"; $GOOGLEfreq = "monthly";}	// past 60 days
if ($mod > mktime(0,0,0) - 86400*30)	{$GOOGLEpriority = "0.7"; $GOOGLEfreq = "monthly";}	// past 30 days
if ($mod > mktime(0,0,0) - 86400*7)	{$GOOGLEpriority = "0.8"; $GOOGLEfreq = "weekly";}	// past 7 days
if ($mod > mktime(0,0,0) - 86400)	{$GOOGLEpriority = "0.9"; $GOOGLEfreq = "daily";}	// yesterday
if ($GOOGLEmoddate == date("Y-m-d"))	{$GOOGLEpriority = "1.0"; $GOOGLEfreq = "hourly";}	// today

IDIF is a stupid format: it includes the entire blog_contents, so the files are huge. In the process of setting this up, I learned that flat files have a maximum size of 1.4 megs or so (the size of an old floppy disk), so I had to create more than one.

Which explains the stupid concept of a "pointer file"; instead of just giving Yahoo the IDIF file itself, you give it a pointer file with URLs pointing to the multitude of IDIF files. Really stupid.

News flash, after finding the Journal Of Ovid on the web, I learned about length restrictions for the input fields (described below). This information was not contained on the Yahoo web site describing their file format. It considerably reduced the file sizes but I retained the structure of multiple files because who knows what I'll learn next?

IDIF title must be a maximum of 80 characters
IDIF description must be a maximum of 180 characters
IDIF body must be a maximum of 1000 characters
I'm only guessing about keywords

Thanks to the Journal Of Ovid on the web for this secret information

From the inside out: trim, replace whitespace (thanks to the PHP manual for this), shorten to maximum length

$IDIFtitle		= substr( preg_replace ('/\s\s+/', ' ', trim($title) ), 0, 80 );
$IDIFdescription	= substr( preg_replace ('/\s\s+/', ' ', trim($description) ), 0, 180 );
$IDIFkeywords		= substr( preg_replace ('/\s\s+/', ' ', trim($keywords) ), 0, 79 ) . " ";
$IDIFblog_contents	= substr( preg_replace ('/\s\s+/', ' ', trim($blog_contents) ), 0, 1000 );

Yahoo is said to support a simple text file list of URLs "urllist.txt" Documentation, of course, is scarce

(my thanks to for HTML encoding)

Web hosting providers

Choosing a web hosting service provider is difficult. There's no brand to rely on and an Internet search turns up confusing claims and offers. We have used two providers:

  • Internet Planners is the service currently running this web site. They have provided what they promised. They do not run the latest versions of MySQL or PHP which means certain newer features are not available but it probably improves their stability and what they offer is, in fact, provided. What was missing that we noticed included
    • RSS parsing (Magpie was a good substitute)
    • and HTML Tidy, which we've just lived without, substituting our own Regular Expressions.

  • Network Solutions, on the other hand, has provided very poor service and has high fees. They run later versions of PHP and MySQL than Internet Planners, but their implementation is poor and the result is that critical functions are not available: user authentication for example (how can an Internet service provider not support user authentication?).

Based on our experience, Internet Planners is a reasonable choice for web hosting; Network Solutions is a bad choice.

Server-Side gzip Compression

Compression can reduce the size of the text (not images) of your web pages as they are transmitted outbound to the client. This will have only a small impact on response time over modern fiber connections but it will significantly reduce your bandwidth consumption (70% on average on this site.)

In XHTML vs. HTML I show how I implemented gzip compression on this site. The problem with that method is that it's a pain. So on another website I tried out the Apache htaccess method to instruct the server to compress all outbound pages. Works like a charm.

# See

# Insert filter
SetOutputFilter DEFLATE

# Netscape 4.x has some problems...
BrowserMatch ^Mozilla/4 gzip-only-text/html

# Netscape 4.06-4.08 have some more problems
BrowserMatch ^Mozilla/4\.0[678] no-gzip

# MSIE masquerades as Netscape, but it is fine
# BrowserMatch \bMSIE !no-gzip !gzip-only-text/html

# NOTE: Due to a bug in mod_setenvif up to Apache 2.0.48
# the above regex won't work. You can use the following
# workaround to get the desired effect:
BrowserMatch \bMSI[E] !no-gzip !gzip-only-text/html

# Don't compress images
SetEnvIfNoCase Request_URI \
\.(?:gif|jpe?g|png)$ no-gzip dont-vary

# Make sure proxies don't deliver the wrong content
Header append Vary User-Agent env=!dont-vary

Call KML files from within a blog or topic

Here's how to create a button in your blogs or topics that calls KML or KMZ files you create. In the Modify A Blog utility you can include a KML or KMZ file, but to call it explicity from within the blog, you can create a button as shown here.

  1. Create and save your KML file.
  2. FTP the file to the kml folder in Philadelphia Reflections
  3. Use the following code in a blog or topic to call the kml file

<button onclick="location.href=''">Button Label</button>

To create this button:

Instead of Franklin.kmz, put any .kml or .kmz file that is in the kml folder:

Ternary Operator and the IIF function

The standard PHP If statement can be reduced by the ternary operator, which is described in the PHP manual Comparison Operators. The IIF function puts the ternary operator into a function.

Ternary Operator

conditional ? if_true : if_not_true;

is the same as

if (conditional)

IIF Function

To return the result of the ternary operator

function iif($expression, $returntrue, $returnfalse = '') {
    return ($expression ? $returntrue : $returnfalse);


The panel below shows every image (2000+) in every blog (800+) on Philadelphia Reflections starting with most recent additions.

It works better on some browsers especially Firefox than others, and -- with 2000 images -- it takes a while to load, as much at 10 minutes on a slow connection. An icon in the corner of the picture-wall starts a slideshow. Note: Mouse-clicking enlarges each thumbnail picture, displaying an icon linked to the website source page. We suggest you try out every little icon to see the amazing versatility of Cooliris.

Embed Flash as Valid XHTML

The problem to be solved is that you want to embed YouTube (or other Flash movies) but the <embed> tag is deprecated in XHTML and the <param> tags don't validate, either. Here are the steps to clean things up:

The original HTML from YouTube


Step 1: replace the closing "</param>" tags with trailing " />"

  value=""  />
  value="true"  />
  value="always"  />

Step 2: put type="application/x-shockwave-flash" into the <object> tag:

  value="" />
  value="true" />
  value="always" />

Step 3: move src="..." from the <embed> tag to a data="..." attribute in the <object> tag:

  value="" />
  value="true" />
  value="always" />

Step 4: remove the <embed> tag:

  value=";hl=en&amp;fs=1" />
  value="true" />
  value="always" />

You can View > Source to see that the code shown here does actually produce the YouTube video displayed.

Step X: it should be noted that in Firefox you don't need any "<param>" tags at all, which makes things very simple and clean:


Not in IE, though; nope. (Why the </object> instead of a closing " />"? Because it seems to work more reliably in Firefox 3.0.5; I don't know why.)

Stylesheet for ATOM feed

Even though neither IE 7 nor Firefox 3.0.8 will render a stylesheet for an ATOM or RSS feed delivered over the network, we have set them up in hopes that someday this anomaly will be cured.

No aggregator I have ever seen goes to the basic trouble of sorting their input feeds by modified date, relying on the feed creator to "push" the latest onto the top of the stack. Our XSL stylesheet does sort by modified date, among other nice things.

Here's the XSL stylesheet for our ATOM feed, including sorting the entries:

<?xml version="1.0" encoding="utf-8"?>

<!--                                   -->
<!--     Philadelphia Reflections      -->
<!--   XSL Stylesheet for ATOM feed    -->
<!--                                   -->

<!-- Grateful acknowledgement to -->

<xsl:stylesheet version="1.0"
    <xsl:output method="html" encoding="utf-8"/>
    <xsl:template match="/">
      <title>ATOM Feed for Philadelphia Reflections</title>
      <link rel="stylesheet" href="" type="text/css"/>
      <link rel="stylesheet" href="" type="text/css"/>
      <style type="text/css">
        .notvisible {visibility: hidden;}
        <xsl:apply-templates select="/atom:feed"/>
    <xsl:template match="/atom:feed">
      <div class="topbox">
        <p><img src="" alt="ATOM feed icon" /> 
        This is the <strong>ATOM-feed </strong> 
        for the <a href=""><xsl:value-of select="atom:title"/></a>
        website.<br />
        ATOM feeds allow you to stay up to date with the latest additions and changes 
        on  <xsl:value-of select="atom:title"/>.</p>
      <div class="contbox">
            <div class="mainbox">
            <div class="itembox">
              <h1><xsl:value-of select="atom:title"/></h1>
              <p><xsl:value-of select="atom:subtitle"/></p>

              <ul id="entries">
                <xsl:apply-templates select="atom:entry">
                  <xsl:sort select="atom:updated" order="descending"/>


          <td valign="top" width="30%">
            <div class="subscrbox">
            <div class="padrhsbox">
              <h2>Subscribe to this feed</h2>
              <p>If you use one of the following web-based News Readers,
                click on the appropriate button to subscribe to the RSS feed.</p>
              <a href="#" onClick="window.location='' + window.location;return false;">
                <img height="17" width="91" vspace="3" border="0" alt="my yahoo" src=""/>
              <a href="#" onClick="window.location=''+ window.location;return false;">
                <img height="18" width="91" vspace="3" border="0" alt="bloglines" src=""/>
              <a href="#" onClick="window.location=''+ window.location;return false;">
                <img height="17" width="91" vspace="3" border="0" alt="newsgator" src=""/>
              <a href="#" onClick="window.location=';a=' + window.location + '&amp;t={title}';return false;">
                <img src="" vspace="3" border="0" alt="Subscribe with Pluck RSS reader"/>
              <a href="#" onClick="window.location='' + window.location;return false;">
                <img src="" vspace="3" border="0" alt="Subscribe in Rojo"/>
              <a href="#" onClick="window.location='' + window.location;return false;">
                <img src="" vspace="3" border="0" alt="Add to Google"/>
              <hr />
              <p>If you would like to receive an email whenever changes are made, please send me an email and I'll be glad to add you. 
              <br /><br /><a href="">Click to send an email</a></p>


    <xsl:template match="atom:entry">
      <li style="margin-bottom: 25px; height: auto;">
        <a href="{atom:link/@href}">
          <xsl:value-of select="atom:title"/>
          &#8212; <xsl:value-of select="substring-before(atom:updated,'T')"/>
        <div class="item_desc">
          <xsl:value-of select="atom:summary" disable-output-escaping="yes"/> <!-- disable-output-escaping="yes" does not work with Firefox 3.0.8 -->


SQL To Exclude A List Of Items

Let's say you have a table "Primary" that contains an "Email" field.

You would like to select all the email addresses in Primary except for the list of email addresses in the Email field in a table "Exclude".

This SQL will exclude the emails in "Primary" based on those contained in "Exclude".

SELECT * FROM Primary WHERE ((Primary.Email) Not In (SELECT Email FROM Exclude))

MySQL server has gone away

MySQL timeout? Probably a new error after years of working perfectly, resulting from an ISP change which they will neither acknowledge nor fix. Sound familiar?

Charming people.

Try this:

$db_link = @mysql_connect(DB_HOST, DB_USER, DB_PSWD,'',MYSQL_CLIENT_INTERACTIVE);

... instead of what you used to do:

$db_link = @mysql_connect(DB_HOST, DB_USER, DB_PSWD);


Function to make inserting new rows into a database table easier (and safe because quote_smart logic is included inline)

thanks to R. Bradley @; I have fixed a number of bugs and added quote_smart functionality

My own contribution to is here: george at georgefisher dot com

function mysql_insert_assoc ($my_table, $my_array) {
// Insert values into a MySQL database
// Includes quote_smart code to foil SQL Injection
// A call to this function of:
//  $val1 = "foobar";
//  $val2 = 495;
//  mysql_insert_assoc("tablename", array(col1=>$val1, col2=>$val2, col3=>"val3", col4=>720));
// Sends the following query:
//  INSERT INTO tablename (col1, col2, col3, col4) values ('foobar', 495, 'val3', 720)
    global $db_link;
    // Find all the keys (column names) from the array $my_array
    $columns = array_keys($my_array);

    // Find all the values from the array $my_array
    $values = array_values($my_array);
    // quote_smart the values
    $values_number = count($values);
    for ($i = 0; $i < $values_number; $i++)
      $value = $values[$i];
      if (get_magic_quotes_gpc()) { $value = stripslashes($value); }
      if (!is_numeric($value))    { $value = "'" . mysql_real_escape_string($value, $db_link) . "'"; }
      $values[$i] = $value;
    // Compose the query
    $sql = "INSERT INTO $my_table ";

    // create comma-separated string of column names, enclosed in parentheses
    $sql .= "(" . implode(", ", $columns) . ")";
    $sql .= " values ";

    // create comma-separated string of values, enclosed in parentheses
    $sql .= "(" . implode(", ", $values) . ")";
    $result = @mysql_query ($sql) 
              OR die ("<br />\n<span style=\"color:red\">Query: $sql UNsuccessful :</span> " . mysql_error() . "\n<br />");

    return ($result) ? true : false;

mysql_update_assoc is a similar function that updates existing records.

Also thanks to for encoding


Function to make updating rows in a database table easier (and safe: quote_smart logic is implented inline).

function mysql_update_assoc ($my_table, $my_array, $where_conditions) {

// Update values in a MySQL database table
// Includes quote_smart code to foil SQL Injection
// A call to this function of:
//  $val1 = "foobar";
//  $val2 = 495;
//  mysql_update_assoc("tablename", array(col1=>$val1, col2=>$val2), array(table_key=>52, age=>"old"));
// Sends the following query:
//  UPDATE tablename SET col1 = 'foobar', col2 = 495 WHERE table_key = 52 AND age = 'old'
//                  -- and --
//  $table_name = "tablename";
//  mysql_update_assoc($table_name, array(col1=>$val1, col2=>$val2), array(table_key=>52));
// Sends this:
//  UPDATE tablename SET col1 = 'foobar', col2 = 495 WHERE table_key = 52
// Note: the WHERE clause is always "=" and always AND

global $db_link;

$sql = "UPDATE $my_table SET ";

// quote_smart the data values and create a comma-separated string of column_name = value
foreach ($my_array as $key => $value)
  if (get_magic_quotes_gpc()) { $value = stripslashes($value); }
  if (!is_numeric($value))    { $value = "'" . mysql_real_escape_string($value, $db_link) . "'"; }
  $sql .= "$key = $value, ";
$sql = substr($sql, 0, -2);  // remove trailing ", "

// quote_smart the conditional values and create a comma-separated string of column_name = value AND
$conditional_pairs = NULL;
foreach ($where_conditions as $key => $value)
  if (get_magic_quotes_gpc()) { $value = stripslashes($value); }
  if (!is_numeric($value))    { $value = "'" . mysql_real_escape_string($value, $db_link) . "'"; }
  $conditional_pairs .= "$key = $value AND ";
$conditional_pairs = substr($conditional_pairs, 0, -5);  // remove trailing " AND "

$sql .= " WHERE $conditional_pairs";

$result = @mysql_query ($sql) 
          OR die ("<br />\n<span style=\"color:red\">Query: $sql UNsuccessful :</span> " . mysql_error() . "\n<br />");

return ($result) ? true : false;

mysql_insert_assoc is a similar function that adds new records.

Thanks to for encoding

How to detect an iPhone and other mobile devices

Handheld/mobile devices have been exploding in popularity and with the advent of the iPhone they have become the device of choice. The Blackberry was a lovely device but once you try an iPhone you will never want a Blackberry again. Of course, all of this will change as each new device comes out but what won't change is the fact that mobile devices are supplanting PCs for everything but the most keyboard- or large-screen-intensive work.

Therefore, the popularity of a website/blog/whatever depends on making it accessible to mobile devices. Step one is knowing when you've been visited by such a thing.


» Articles

The iPhone is very well-behaved with respect to CSS. Simply include the following meta tag in an otherwise-ordinary web page:

<!--[if !IE]>-->
<link rel="stylesheet" 
  media="only screen and (max-device-width: 480px)" 
  href="iphone.css" />

... which points to the iphone.css CSS stylesheet:

body { margin: 0; 
       padding: 0; 
       width: 100%; }				
    #wrapper { margin: 0; padding: 0; width: 100%; }
      #left    { display: none; }
      #right   { display: none; }
      #center  { margin: 0; }
        #welcome  { display: none; }
        #content  { line-height: 115%; 
                    font-family: "Times New Roman", Times, serif; 
                    padding: 5px 20px 0 20px; 
                    margin: 0;
                    font-size: 28px; 
                    background-color: white; }
        #comments { display: none; }
      #footer  { display: none; }

iphone.css uses { display: none; } to keep the surrounding boxes from displaying on the smaller screen. No other changes are required.

An excellent article on using CSS is here:

(Subsequent to writing this article we wrote an iPhone-specific article page. The stylesheet method described does work but we had other reasons to make modifications to the content.)

» Index Page

The index page was a complete rewrite of the standard index page to turn it into a simple table of contents. The breakthrough here was an exquisite PHP script found at which analyzes the HTTP headers to determine if a device is a handheld and if so, what type. Using this, we redirect from the regular index page to the iPhone-specific index page (from index.php to indexiphone.php ... it looks much better on an iPhone than on a PC).

Generic Handheld Devices

We may build more device-specific CSS files and pages as we learn what our visitors use but for the time being we simply support iPhones and Other.

Because many handheld devices have very small screens and a very small buffer capacity, we strip out all HTML comments, images and tables; we also convert to UTF-8 encoding because of the claim that handhelds support it better than iso-8859-1:

$content = preg_replace('/<!-- .*? -->/si', '', $content);
$content = preg_replace('%<table [^>]*?>.*?</table>%si', '', $content);
$content = preg_replace('/<img [^>]*?>/i', '', $content);
echo utf8_encode($content);

We include these meta tags (cribbed from Google's mobile page):

<meta name="viewport" content="width=device-width,minimum-scale=1.0,maximum-scale=1.0"/>
<meta name="HandheldFriendly" content="true" />

We also send handheld-specific HTTP headers ... see XHTML vs. HTML for the script we use for all page headers.

» Articles

It turns out in real life that many handhelds don't handle CSS correctly, so the { display: none; } trick that works so well for iPhones is unreliable for many other devices. Furthermore, many devices choke if sent a large data stream. Therefore, we had to write a very stripped-down page for each individual article. This required writing only one new program since all of our content is served from a database.

The function is used to redirect from the regular page to the handheld page.


A glance at the page source of each of these pages will show you the extent to which the handheld version is stripped down; rather elegant, really.

Finally, in the regular pages we include this link tag; wishful thinking I suspect:

<link rel="alternate" 
  href=";key=####" />

» Index Page

The generic handheld index page ( is an even-more stripped down version of the iPhone index page (

Testing and Validating

The place to start testing your new pages is the Opera Mini Simulator: Elegant in its simplicity, it has never choked or failed, unlike most other emulators. Plus, it is completely intuitive; unlike all other emulators. And free.

Once you get the basic design underway, you will want to validate ... at

For fine-tuning of displays on cell phones ...

To run the Blackberry emulators on a Windows machine:

I found help for this ridiculous process at You may, too.

Only a sadistic socially-crippled geek-savant could have dreamed up such a convoluted mess, but ultimately it does work and does allow you to see how the different models operate. Actually, once you get it working,it's sort of fun to try different models.

(my thanks to for HTML encoding)

An iPhone web app

The iPhone is the best PDA to come along since the Blackberry 15 years ago. It is to the Blackberry what the Blackberry was to cell phones.

Philadelphia Reflections is now a fully-fledged iPhone web app. The application will appear on your iPhone in the appropriate format automatically: just navigate to with the iPhone browser; we will detect it and do the right thing.

A two-step process is required to get a little icon on your iPhone home page so you can go there directly:

  1. Click the "+" plus sign at the bottom of the iPhone screen
  2. Click the "Add to Home Screen" button that appears.

That's it. We do all the rest.

We are listed in the Apple web-app store: click here to go see our special page.
Click Add to Home Screen
Click the plus sign

Valid XHTML YouTube embed code generator

Escaping for PHP Output to JavaScript

To send data to a JavaScript script from PHP, three levels of escaping are required as shown in the snippet below:

<script type="text/javascript">
// <![CDATA[

// escape for JavaScript
$message   = preg_replace("/\r?\n/", "\\n", addslashes($message));

// escape for XHTML
$message   = preg_replace('%</%i', '<\/', $message);

// send to JavaScript
echo "   var message = \"$message\";\n";

// ]]>

(my thanks to for HTML encoding)

Google Maps Icons

Google Maps/Earth do not make icon creation & manipulation easy. Here are a couple of tips:

GIcon (look here: has a number of methods for creating and modifying an icon. I've found it's best to start with the default because adding features you expect is harder than you think.

// Here's how to create a new icon with the defaults
var newIcon = new GIcon(G_DEFAULT_ICON);

// To create a new icon like the default but yellow:
var yellowIcon = new GIcon(G_DEFAULT_ICON, "");

// To make use of that yellow icon and give it a tooltip
var point   = new GLatLng(40.39, -75.34);
var marker  = new GMarker(point, {icon:yellowIcon, title:"View Above Philadelphia"});

// To open a balloon when clicked
var message = " ... fill with text and HTML ... I've found tables are very helpful ";
GEvent.addListener(marker, 'click', function() {marker.openInfoWindowHtml(message);});

// Change icon on mouseover (see
var msoverIcon = new GIcon(G_DEFAULT_ICON, "");
GEvent.addListener(marker, 'mouseover', function() { marker.seticon(msoverIcon); });

Note: the "message" part of that snippet has to be escaped correctly.

(my thanks to for HTML encoding)

ZNOTE: Website Development


Display Image

This clever website will create an image with the specified dimensions

<img src="" alt="A Dummy Image" />

{A Dummy Image}
Image created with specific dimensions

Python URL Handling

In case you're wondering "How the heck does Python handle headers and data under 3.2.2?", here's an example that works using IDLE and Python 3.2.2 installed on a 64-bit Windows 7 machine.

import re
import urllib.request
url = ""

uf = urllib.request.urlopen(url)

# header information
print('--- headers ---')
info =  # headers

#headers = info._headers # a list of all the headers


print('--- data ---')

data = # content

print('--- image ---')

imageurl = url + "/images/001.JPG"
image = urllib.request.urlretrieve(imageurl, 'python_001.jpg')

(my thanks to for HTML encoding)

QR Codes

QR Codes are similar to bar codes in that they are read optically. Most-common in Japan, all Japanese cell phones can read them; all fancy phones in America (iPhone, etc.) have download-able apps that can read QR Codes (semacode is a free QR Code app for the iPhone but there are many for all).

One application that is becoming common is encoding a website's URL and including the image in a print advertisement.

The QR Codes below were created by; create QR and semacode/DataMatrix from text:

QR Codes
{philadelphia reflections qrcode} {george fisher qr code}
Philadelphia Reflections George Fisher (Flash... N/G on iPhone)
{chemical heritage society qr code} {kaiser qr code}
Chemical Heritage Society Kaiser Permanente
{george fisher advisors qr code}
George Fisher Advisors QR Code George Fisher Advisors semacode/DataMatrix

Website Test Results

The site is an excellent facility for testing the performance of a web site. Philadelphia Reflections passes with flying colors:

  • A First Byte Time ... the PHP program is efficient
  • A Keep-alive Enabled ... the TCP connection is kept open throughout the entire process
  • A Compress Text ... the text is gzip compressed to reduce the amount of data sent
  • A Compress Images ... we reduce the size of our images and use JPEG compression
  • A Cache Static Content ... this can be seen in the Repeat View: none of the images are sent on subsequent page loads
  • X CDN Detected ... large commercial sites have copies scattered around the country and the world to balance the load and to reduce latency but it would not pay for a small site like this one to go to this effort and expense
{Response time test with lots of images}
Response time test with lots of images


Excellent site for website testing

Iterate through a Word document, modifying picture properties(Blog 2300)

(Blog 2300) We have a facility on this website to download books of many chapters (made up of volumes of topics on the site) to Microsoft Word for subsequent editing and eventual publishing. In many cases we download lots of pictures (via an img src= tag). I have not found a way to set the way text flows around the images in Word using HTML or CSS, so I built a Word macro to do it. This should allow you to change the size of images, as well as move them around. Moving the captions requires the use of the captions feature in Word's image menu (right-click).


Instructions for use of a Macro named Sub ImageFlow():

  1. Open Word

  2. In Word, enter File>Open

  3. enter the URL of the file you want to modify into the File Entry box and press the Enter key to load the document. It may take a minute or two, but a working screen should appear, loaded with the file in a condition ready to move the pictures around.

  4. Press Alt + F11 which will open the VBA screen

  5. Copy the macro found on this page from
    Sub ImageFlow()
    End Sub
  6. In the right-hand panel of the VBA screen press Ctrl+A, Ctrl+V to paste it in

  7. In the VBA screen press F5 to run the macro

If you want to do a lot of these manipulations, save the macro in the Macro Library of Windows Word.

Sub ImageFlow()
'  this Macro goes through an entire Word document and
'  changes the way text flows around each picture
'  ("Tight" in this example but see below for choices)
    Dim shpIn As InlineShape, shp As Shape

    For Each shpIn In ActiveDocument.InlineShapes
        If (shpIn.Type = wdInlineShapeLinkedPicture) Then
            Set shp = shpIn.ConvertToShape
            shp.WrapFormat.Type = wdWrapTight
        End If
    Next shpIn

    For Each shp In ActiveDocument.Shapes
        shp.WrapFormat.Type = wdWrapTight
    Next shp

End Sub

Change wdWrapTight to any of the following:

My thanks to for showing me the essential technique of iterating through the pictures.

What are the InlineShapes' Types? See; it is possible we may also need to select on wdInlineShapePicture (as well as wdInlineShapeLinkedPicture) but for my specific purpose I did not need to.

SMTP Authorization and Handling Bounced Emails with PEAR Mail

Recently our ISP started requiring user signon in order to send emails. PHP's mail function stopped working as a result.

Naturally, the ISP did not notify us of this change so we were quite surprised when many thousands of emails on our newsletter list were rejected (every one of them, in fact).

What error message was returned to us to notify us of what the problem was? Why this helpful note:

Mail sent by user nobody being discarded due to sender restrictions in WHM->Tweak Settings

Doesn't that just say it all?

I'm being snide, but our ISP is really quite good about keeping its software up to date and aside from an occasional surprise like this, they are very reliable. Being up to date included the automatic incorporation of the PEAR Mail facility which we are now using.

PEAR's Mail system works quite well but two problems were very vexing until we stumbled our way to a solution:

  1. How, exactly, do we sign on to the SMTP server?
  2. How do we ensure that bounced emails (the bane of all email lists) get returned to us?

You might not think that the first question would be so hard but it actually took a good deal of trial and error to get it right. As for the second question, there is an awful lot of wrong information available out in Internet land (including but not limited to VERP and XVERP which I advise you to avoid).

With PEAR Mail you first set up a "factory" and then send emails, either singly or in a loop. We keep the user id, password, etc. in a file "above" the web server in hopes that will keep them secret ... here's the code (it actually is in production and it does in fact work):


# the email constants are contained in a file outside the web server

$headers = array (
         'From' => '"name"<>',
         'Sender' => '"name"<>',
         'Reply-To' => '"name"<>',
         'Return-Path' => '',
         'Content-type' => 'text/html; charset=iso-8859-1',
         'X-Mailer' => 'PHP/' . phpversion(),
         'Date' => date("D, j M Y H:i:s O",time()),
         'Content-Language' => 'en-us',
         'MIME-Version' => '1.0'

// call the PEAR mail "factory"
$smtp = Mail::factory('smtp',
      array (
            'host' => EMAIL_HOST,
            'port' => EMAIL_PORT,
            'auth' => true,
            'username' => EMAIL_USERNAME,
            'password' => EMAIL_PASSWORD,
            'persist' => true,
            'debug' => false
            ), '-f'

# to send emails:
# $headers['To']      = $to;        # provide the "$to" variable, something like $to = '"name"<>';
#                                   # note that the first parameter of $smtp->send can be "decorated" this way or just a naked email address
# $headers['Subject'] = $subject;   # provide the "$subject" variable
# $mail = $smtp->send($to, $headers, $contents_of_the_email);
#                          -------- ................................> except for 'To' and 'Subject',
#                                                                     $headers is provided by this module but can be over-ridden
# if (PEAR::isError($mail))
# {
#   echo "<p style='color:red;'>The email failed; debug information follows:<br />";
#   echo $mail->getDebugInfo() . "<br />";
#   echo $mail->getMessage()   . "</p>";
# }
# else
# {
#   echo "<p>email successfully sent</p>";
# }


My thanks to for the HTML entites conversion.

When can you start?
Posted by: Eipmwlod   |   Jan 25, 2012 3:55 PM
I have not been to this blog for a long time, however it

was a joy to find it again. It is such an important

topic and ignored by so many, even professionals! I

thank you for helping to make people more aware of these

issues. Just great stuff as per usual.For more

information go to <>.
Posted by: Johnmaurice   |   Sep 26, 2011 6:40 AM
this post is fantastic %3
Posted by: Yutqlnvp   |   Feb 5, 2011 11:09 PM
Some pretty amazingly esoteric information here, mate!
Posted by: Late Nights   |   Mar 16, 2007 10:06 PM
Good work. Interesting posts, besides those spam...
Posted by: Oopa Jopa   |   Oct 11, 2006 6:01 PM
C'est trouis bien. Nice, i mean. Thanks!
Posted by: Junior Lee   |   Sep 26, 2006 11:07 AM

Please Let Us Know What You Think

(HTML tags provide better formatting)

Because of robot spam we ask you to confirm your comment: we will send you an email containing a link to click. We apologize for this inconvenience but this ensures the quality of the comments. (Your email will not be displayed.)
Thank you.